3D Vision
3D spatial perception technology that processes data from structured-light, stereo and ToF 3D cameras to analyze the position, size and pose of irregular objects.
It delivers reliable recognition even with low-cost ToF 3D cameras, and can be applied to automation in logistics and a wide range of other industrial sites with irregular items.
3D Camera-Based Spatial Perception
3D spatial perception technology that lets robots understand their work environment and analyze the position and shape of objects.
It supports a variety of 3D cameras, including ToF, structured light and stereo,
and is optimized for low-cost ToF cameras, lowering deployment costs and making automation more practical at industrial sites with irregular items.
Key Features of 3D Vision
013D Object Parameter Extraction
This technology estimates physical properties of an object, such as its center coordinates, size and pose, from 3D data. Using deep learning for 3D spatial feature extraction and spatial segmentation algorithms, we are conducting research to analyze the shape of even partially occluded objects. It can be applied to recognize a wide range of irregular objects, including logistics boxes, parts and materials.
02Automation with Low-Cost Sensors
It supports a variety of 3D camera setups, including ToF, structured light and stereo, and a low-cost ToF camera architecture in particular eases the cost of adopting automation. Reliable object recognition across logistics, manufacturing, picking and other industrial sites makes it easier to deploy in the field.
03Adjacent Object Separation
Developed to take advantage of the cost efficiency of ToF cameras while compensating for their relatively limited resolution. It separates objects that sit close together or partially overlap, enabling reliable recognition of irregular objects such as logistics boxes.
Key Specifications
Detailed specifications for technical review.
| 3D cameras | Supports a variety of 3D cameras, including ToF, structured light and stereo |
|---|---|
| Recognized data | Object picking pose, position and size |
| Recognition method | Object analysis based on 3D data |
| Key strength | Reliable object recognition even with low-cost ToF-based 3D cameras |
| Adjacent objects | Individual identification and classification of adjacent or partially overlapping objects |
| Occluded objects | Research in progress on estimating the shape of partially occluded objects |
| Competitive edge | Cost-optimized solutions matched to required precision and budget |
| Application | ToF for tasks where high precision and cost efficiency both matter; structured light for ultra-high-precision tasks |
Ongoing Collaboration
Our 3D spatial perception technology is being advanced with government R&D support.
Selected for 2025 Deep Tech TIPS
Selected for Korea's Ministry of SMEs and Startups' Deep Tech TIPS program, we are using approximately KRW 2 billion in R&D funding to advance RBS 2.0, an unmanned logistics automation system that incorporates our 3D perception technology.
Looking for the right product or solution?
Request a Demo & Site Assessment